AGESA 1.0.0.6b Might Fix The Ryzen Linux Performance Marginality Problem

[LinuxID10T]

That bios doesn't have the agesa 1.0.0.6b according to reddit.

Dave.

I stand corrected then. It was just perfectly timed and the only comment to the bios update was "Improve System stability" therefore I assumed it was supposed to solve this problem.

[eydee]

I stand corrected then. It was just perfectly timed and the only comment to the bios update was "Improve System stability" therefore I assumed it was supposed to solve this problem.

It was a patch to fix the memtest86 issue. 1.0.0.6b isn't out on any Asus boards yet.

[L_A_G]

I stand corrected then. It was just perfectly timed and the only comment to the bios update was "Improve System stability" therefore I assumed it was supposed to solve this problem.

When they change the underlying AGESA version it's always mentioned in the patch notes of a BIOS update.

I however get the feeling you're far from the only one who's updated to a BIOS without 1.0.0.6b (MSI doesn't have any ones with it either and plenty of Gigabyte boards also lack it) and complained that it doesn't fix it...

[SaucyJack]

Without doing any heavy testing, it looks like a recent kernel update might have fixed it. The last two times I compiled mesa I didn't get a single segfault. I ALWAYS got at least one segfault compiling mesa before.

[Chewi]

Ha and I'm actually going to be packing up my CPU to send back this evening! I was just waiting for some thermal grease remover to arrive. I progressed through the RMA process without any issues in the space of about a week.

[curaga]

With Linux able to update microcode itself, why would you need a bios update? Or is AMD acting like a jerk, and only giving the microcode update to mobo vendors for now?

[SaucyJack]

nevermind, 21 minutes to fail with kill ryzen script.

[milhouse]

With Linux able to update microcode itself, why would you need a bios update? Or is AMD acting like a jerk, and only giving the microcode update to mobo vendors for now?

Agreed. It would be nice if AMD provided updated microcrode for fam17h (Zen) as there's nothing here so far: https://git.kernel.org/pub/scm/linux...tree/amd-ucode

Not everyone wants to update their BIOS to fix CPU bugs that can be taken care of by the OS. And distributions can include the latest CPU microcode and have the peace of mind it brings now that they don't have to worry about contending with a buggy CPU due to out of date BIOS. Everybody wins. So come on AMD...

[Azrael5]

I've just red a post where the problem is addressed to some Ryzen CPus realized before 25th weeks of the year. Is it confirmed?

This is the post: https://community.amd.com/message/2821277

[carewolf]

With Linux able to update microcode itself, why would you need a bios update? Or is AMD acting like a jerk, and only giving the microcode update to mobo vendors for now?

Linux would overwrite the microcode then. I am guessing it is changing some voltages or hardware timings instead.